從燒瓶請求Python獲取空

Question

我正在編寫一個簡單的flask應用程序，根據我的查詢，我應該以所需的格式獲得所需的答案。 代碼如下：

#-*- coding: utf-8 -*-
import StringIO
import os
import pandas as pd
import numpy as np
from flask import Flask, request, Response, abort, jsonify, send_from_directory,make_response
import io
from pandas import DataFrame
import urllib2, json
import requests
from flask import session
import sys  


reload(sys)  
sys.setdefaultencoding("ISO-8859-1")
app = Flask(__name__)

@app.route("/api/conversation/", methods=['POST'])
def chatbot():
    df = pd.DataFrame(json.load(urllib2.urlopen('http://192.168.21.245/sixthsensedata/server/Test_new.json')))
    question = request.form.get('question')
    store = []

    if question == 'What is the number of total observation of the dataset':
       store.append(df.shape)

    if question == 'What are the column names of the dataset':
       store.append(df.columns)
    return jsonify(store)

if __name__ == '__main__':
    app.debug = True
    app.run(host = '192.168.21.11',port=5000)

它運行正常，但響應為空。 我想再創建30個以上的問題，並將值存儲在store數組中。 但是，我認為價值並沒有被附加到store內部。

不過，在jupyter筆記本中，我得到了適當的答復；

df = pd.DataFrame(json.load(urllib2.urlopen('http://192.168.21.245/sixthsensedata/server/Test_new.json')))
store = []
store.append(df.shape)
print store
[(521, 24)]

為什么在flask中，值未附加？ 我正在郵遞員中測試我的應用程序。 請指導我缺少的地方。

郵遞員的屏幕截圖

Answer 1

如果不提供Post方法的數據類型，則request.form的計算結果為

ImmutableMultiDict([('{"question": "What is the number of total observation of the dataset"}', u'')])

和question = request.form.get('question')最終都不是。您可以顯式使用內容類型作為json，或強制加載它。

@app.route('/api/conversation/', methods=['POST'])
def chatbot():
    question = request.get_json(force=True).get('question')
    store = []

    if question == 'What is the number of total observation of the dataset':
        store.append("shape")
    elif question == 'What are the column names of the dataset':
        store.append("columns")
    return jsonify(store)

卷曲請求

$curl -X POST -d '{"question": "What is the number of total observation of the dataset"}' http://127.0.0.1:5000/api/conversation/

[“形狀”]

$curl -H 'Content-Type: application/json' -X POST -d '{"question": "What is the number of total observation of the dataset"}' http://127.0.0.1:5000/api/conversation/

[“形狀”]